This paper presents an estimation of the rotor power coefficient (C p) curve, which is useful for pitch angle control of a wind turbine system. The Cp curve is affected by several factors such as the structure of a wind turbine, surrounding environment in which the wind turbine built, and its method of control, etc. Therefore, it is necessary to estimate this curve in real-time using direct measurements from the generator and wind turbine. To achieve the optimal estimation for the Cp curve, the reduced multivariate polynomial (RMP) model is applied because it can be basically represented in a polynomial form. Unlike general neural network algorithms, the RMP model avoids a training process. This characteristic makes it possible to apply to the real-time estimation in a practical situation. Also, the first-order partial derivatives of the Cp curve are easily computed by using the RMP model. This derivative information can be effectively used to maximize turbine output power by a proper pitch angle control. The simulation results show that the proposed RMP model provides a good estimation performance in a fast and effective manner.